1. Field of the Invention
This invention relates generally to machines and procedures for printing text or graphics on printing media such as paper, transparency stock, or other glossy media; and more particularly to a scanning thermal-inkjet machine and method that construct text or images from individual ink spots created on a printing medium, in a two-dimensional pixel array. The invention employs print-mode techniques to optimize image quality (primarily on transparent and glossy media) vs. operating time, and also to minimize image distortion (primarily on paper) imposed by an ink-drying heater.
2. Prior Art
To achieve vivid colors in inkjet printing with aqueous inks, and to substantially fill the white space between addressable pixel locations, ample quantities of ink must be deposited. Doing so, however, requires subsequent removal of the water base--by evaporation (and, for some printing media, absorption)--and this drying step can be unduly time consuming.
In addition, if a large amount of ink is put down all at substantially the same time, within each section of an image, related adverse bulk-colorant effects arise: so-called "bleed" of one color into another (particularly noticeable at color boundaries that should be sharp), "blocking" or offset of colorant in one printed image onto the back of an adjacent sheet with consequent sticking of the two sheets together (or of one sheet to pieces of the apparatus or to slipcovers used to protect the imaged sheet), and "cockle" or puckering of the printing medium. Various techniques are known for use together to moderate these adverse drying-time effects and bulk- or gross-colorant effects.
(a) Prior heat-application techniques--Among these techniques is heating the inked medium to accelerate evaporation of the water base or carrier. Heating, however, has limitations of its own; and in turn creates other difficulties due to heat-induced deformation of the printing medium.
Glossy stock warps severely in response to heat, and transparencies too can tolerate somewhat less heating than ordinary paper. Accordingly, heating has provided only limited improvement of drying characteristics for these plastic media.
As to paper, the application of heat and ink causes dimensional changes that affect the quality of the image or graphic. Specifically, it has been found preferable to precondition the paper by application of heat before contact of the ink; if preheating is not provided, so-called "end-of-page handoff" quality defects occur--this defect takes the form of a straight image-discontinuity band formed across the bottom of each page when the page bottom is released.
Preheating, however, causes loss of moisture content and resultant shrinking of the paper fibers. To maintain the paper dimensions under these circumstances the paper is held in tension by a system of pinchwheels used in conjunction with paper-advance drivewheels.
Unfortunately these provisions have their maximum effect, in preventing image-quality defects, only while the paper is constrained by the wheels. As soon as the bottom of the page has been printed and the paper leaves the constraint of the wheels, the paper contracts.
This happens very quickly, and as it does the paper and the dots of ink on it move in at the edges and up in the center. The quality defect caused by this sudden releasing of stress can be identified as an "end-of-page paper-shrink defect"; it appears as a thin arched gap of reduced color density.
Prior efforts to eliminate this arched gap have included avoiding the page-long accumulation of stress by cyclically lifting or releasing the constraining force of the pinchwheels. This works to decrease the paper-shrink defect by allowing the internal stress to be released or equalized incrementally--rather than cumulatively.
Unfortunately, however, this cyclical-release technique sacrifices control over paper position at each of the release points along the way. This loss of paper-position control can create numerous misalignment regions that are a greater problem than the paper-shrink defect.
(b) Prior print-mode techniques--Another useful technique is laying down in each pass of the pen only a fraction of the total ink required in each section of the image--so that any areas left white in each pass are filled in by one or more later passes. This tends to control bleed, blocking and cockle by reducing the amount of liquid that is all on the page at any given time, and also may facilitate shortening of drying time.
The specific partial-inking pattern employed in each pass, and the way in which these different patterns add up to a single fully inked image, is known as a "print mode". Heretofore three-pass print modes have been used successfully to reduce bulk-colorant problems on paper--but less successfully on glossy and transparency stock, which are much less absorbent and so rely to a greater extent upon evaporation.
Attempts have also been made to use print modes for hiding the paper-shrink error discussed in subsection (a) above. Heretofore such efforts have had relatively little effectiveness, or have caused still other problems.
For example, some print modes such as square or rectangular checkerboard-like patterns tend to create objectionable moire effects when frequencies, harmonics etc. generated within the patterns are close to the frequencies or harmonics of interacting subsystems. Such interfering frequencies may arise in dithering subsystems sometimes used to help control the paper advance or the pen speed.
Checkerboard print-mode patterns also are subject to objectionable so-called "banding"--horizontal stripes across the finished image. These arise because between each swath the paper advances by substantially the full height of a swath, in effect another type of cumulative-error display.
Print-mode patterns that are instead made up of either mostly all horizontal or mostly all vertical elements can still produce similar interference effects, but only along that direction of the pattern (the direction along which most of the pattern elements are aligned)--and also tend to exaggerate other print-quality defects in the directional lateral to the pattern. Such problems have defeated earlier efforts to find print-mode solutions to the end-of-page paper-shrink problem.
(c) End-of-image print-medium advance errors--Another problem, related to the end-of-page defect introduced above, arises when printing near the beginning or end of a sheet of printing medium--but arises in a somewhat simpler or more mechanical fashion. As suggested earlier, in representative modern printing machines designed for fine resolution and high image quality, the printing medium is generally held taut in the print zone, between two sets of rollers or the like.
This arrangement promotes very high precision and accuracy of printing-medium advance, and thus of printing-medium positioning relative to the pen. Near both longitudinal ends of each sheet or page of printing medium, however, necessarily the medium is held only by one set of rollers etc.
This arrangement leads to relatively less precise positioning of the printing medium in those two regions. This situation may be troublesome in particular when printing near the bottom end of a sheet, as there the sheet is held only by a tensioning roller--which for other reasons is advantageously made rather small in diameter, but such sizing may be adverse to best precision.
One current development (not prior art with respect to the present invention) importantly mitigates that relative diminution of precision by taking smaller steps in the printing-medium advance, particularly near the bottom or end of each page. That system and its benefits are described in the Broder et al. document mentioned earlier; although that system represents a major contribution to attainment of good print quality near the bottom of the page, nevertheless precision is not improved to the level enjoyed in regions where the medium is held taut.
(d) Known technology of print modes: general introduction--One particularly simple way to divide up a desired amount of ink into more than one pen pass is the checkerboard pattern mentioned above: every other pixel location is printed on one pass, and then the blanks are filled in on the next pass.
To avoid the banding problem (and sometimes minimize the moire patterns) discussed above, a print mode may be constructed so that the paper advances between each initial-swath scan of the pen and the corresponding fill-swath scan or scans. In fact this can be done in such a way that each pen scan functions in part as an initial-swath scan (for one portion of the printing medium) and in part as a fill-swath scan.
Once again this technique tends to distribute rather than accumulate print-mechanism error that is impossible or expensive to reduce. The result is to minimize the conspicuousness of--or, in simpler terms, to hide--the error at minimal cost.
For instance a two-pass print mode may start a page by printing with only some of the nozzles in an array of only half of the pen's nozzles, all positioned at one end of the pen--as an example, selected ones of the nozzles consecutively numbered one through fifty, on a hundred-nozzle pen. This first pass may be in a checkerboard pattern--thus actually using, e.g., for example, exclusively odd-numbered nozzles 1, 3, . . . in the first row, and then only even-numbered nozzles 12, 14, . . . in the second row, next selecting only odd-numbered nozzles 21, 23, . . . again in the third row, etc.--and thus printing in half of the pixel locations in the swath area.
The paper then advances by a distance equal to the length of the half-array of nozzles (in other words, the height of fifty nozzles), and the pen would print in both ends of its nozzle array--but again only printing a fifty-percent checkerboard pattern. Now, however, while the forward end of the pen (selected ones of nozzles one through fifty) as before prints on fresh paper, the rearward end (selected ones of nozzles numbered fifty-one through one hundred) fills in the area already printed.
This behavior is then repeated all down the page until the last swath--which is a fill-in swath only, again using selected nozzles of those numbered fifty-one through one hundred.
(e) Space- and sweep-rotated print-mode masks--The pattern used in printing each nozzle section is known as the "print-mode mask". The term "print mode" is more general, usually encompassing a description of a mask, the number of passes required to reach full density and the number of drops per pixel defining "full density".
In the two-pass example above, the second half of the pen (certain ones of nozzles numbered fifty-one through one hundred) filled in the blank spaces left by the first half. For each pass, this may be symbolized using a letter "x" for each pixel that is printed and a letter "o" for each pixel that is not, as follows.
______________________________________ pattern 1: pattern 2: nozzles 1 through 50 nozzles 51 through 100 ______________________________________ xoxoxoxoxo oxoxoxoxox oxoxoxoxox xoxoxoxoxo xoxoxoxoxo oxoxoxoxox oxoxoxoxox xoxoxoxoxo xoxoxoxoxo oxoxoxoxox ______________________________________
In each of these diagrams, the xs appear in diagonal lines--which are angled, if the vertical and horizontal spacings are the same, at forty-five degrees (to both the columns and rows). These lines of xs represent pixels that are printed (if the desired image calls for anything to be printed in each of those pixels respectively), and the os represent diagonal lines of pixels that are not printed.
To conserve space in this document, the diagrams above represent only eight pixel rows, out of fifty created by each half of the hundred-nozzle pen that is under discussion. The nozzles are laid out along the pen in substantially only one vertical row, one hundred nozzles long--although as a practical mechanical matter they are staggered laterally to permit very close spacing along the vertical axis. Therefore to obtain the checkerboard (or other) patterns described in this document the various nozzles are fired selectively and rapidly many times, in careful synchronism with scanning of the pen across the printing medium--taking into account not only the scanning motion across the page but also the nozzle staggering across the pen.
In the "pattern 1" diagram, one line of xs begins in the upper left-hand corner, and at pixel positions offset by two pixels along both top and left-hand edges of the pattern. In the "pattern 2" diagram, however, it is instead a line of as that begins in the corner, whereas lines of xs begin at positions offset from the corner by just one pixel along the top and left-hand edges--and so fitting between the lines of xs put down by "pattern 1".
Hence these diagrams show that pixel positions left unprinted by the first ("pattern 1") pass are filled in by the second. In other words, looking all the way across any row--and taking into account all the xs formed by both "pattern 1" and "pattern 2" in the aggregate--all positions in the row are filled.
One way to achieve this pattern is to always keep nozzles one through fifty in "pattern 1", and always keep nozzles fifty-one through one hundred in "pattern 2". This is known as "space rotated" masking; using this method to print down the page would progressively produce these patterns--illustrated here too using an abbreviated vertical nozzle-array representation of just eight nozzles rather than one hundred:
______________________________________ pass 1 pass 2 pass 3 pass 4 pass 5 ______________________________________ -- .vertline. .vertline. .vertline. .vertline. .vertline. .vertline. xoxo oxox .rarw. first printed row oxox xoxo xoxo oxox oxox xoxo xoxo oxox .rarw. fifth printed row oxox xoxo xoxo oxox oxox xoxo xoxo oxox oxox xoxo xoxo oxox oxox xoxo xoxo oxox oxox xoxo xoxo oxox oxox xoxo .vertline. .vertline. .vertline. .vertline. .vertline. .vertline. -- ______________________________________
In this mode, the pen uses the same pattern all down the page, but the mask is different in different portions of the pen: "pattern 1" for nozzles one through fifty (represented in the abbreviated drawing by the lower four positions in each eight-nozzle group); vs. "pattern 2" for nozzles number fifty-one through one hundred (represented by the upper four positions in each group).
The availability of this method of masking for various printing devices depends in part on the basic mechanical and firmware architecture of each device. In particular, it depends upon whether the basic operating system provides for efficient addressing of different mask patterns to different segments of the overall nozzle array.
Another way to use the same print mode is to apply one mask pattern to the entire pen, but to change that mask pattern from pass to pass. This is so-called "sweep rotated" masking--still using the same abbreviated representation for purposes of illustration:
______________________________________ pass 1 pass 2 pass 3 pass 4 pass 5 ______________________________________ -- .vertline. .vertline. .vertline. .vertline. .vertline. .vertline. xoxo oxox .rarw. first printed row oxox xoxo xoxo oxox oxox xoxo oxox xoxo .rarw. fifth printed row xoxo oxox oxox xoxo xoxo oxox xoxo oxox oxox xoxo xoxo oxox oxox xoxo oxox xoxo xoxo oxox oxox xoxo xoxo oxox .vertline. .vertline. .vertline. .vertline. .vertline. .vertline. -- ______________________________________
In both these diagrams--as in the basic "pattern 1" and "pattern 2" diagrams discussed just before, it can be seen by reading all the way across any row that after both passes at each row all positions in that row are filled--but by comparing the space- and sweep-rotation diagrams it will now be appreciated that the order in which some of the positions are filled in sweep rotation is opposite to that in which they are filled in space rotation. For example, in the fifth printed row the left-hand column is printed in the second pass (and the adjacent column left blank for printing later) in space rotation--but is printed in the third pass (after the adjacent column) in sweep rotation.
This can be shown more compactly by a different notation that allows comparison of space and sweep rotation side by side. In this notation, "0" represents nozzle groups that are not fired at all--at the top and bottom scans of the page--while "1" and "2" represent not individual pixel rows but rather half-swaths, in "pattern 1" and "pattern 2" as defined above.
______________________________________ Space rotation Sweep rotation ______________________________________ 0 0 1 2 1 2 1 2 2 1 1 2 1 2 1 2 2 1 1 2 1 2 0 0 ______________________________________
Now in these abbreviated forms it is easier to see that within the printed image every half-swath receives one "1" and one "2"--but not always in the same order. Thus in the second half-swath the "1" goes down first in space rotation, but second in sweep rotation.
(f) Autorotating print-mode masks--Operating parameters can be selected in such a way that, in effect, rotation occurs even though the pen pattern is consistent over the whole pen array and is never changed between passes. Figuratively speaking this can be regarded as "automatic" rotation or simply "autorotation".
To understand what produces this condition, it is necessary first to take note of what constitutes a basic cell or unit of the print-mode mask, and then to note its height h.sub.c in pixels. It is also necessary to note the number of pixels (or the length measured in number of nozzles) by which the paper moves m.sub.p in each of its advances. For example, in the simple cases diagrammed above, since each mask repeats every two rows, h.sub.c =2; and the paper advances by fifty nozzles at a time, so m.sub.p =50 (or as in the abbreviated-notation diagram the paper advances four diagrammed nozzles at a time, so m.sub.p =4).
The next step is to determine whether the ratio m.sub.p /h.sub.c of these two parameters is integral. If so, as in this case, since m.sub.p /h.sub.c =50/2=25 actually (or 4/2=2 as illustrated), the mask will not autorotate.
If however, in the two-pass example the paper advances by three diagrammed pixel rows instead of four--but the basic cell remains two pixels tall--then for this case as diagrammed the ratio m.sub.p /h.sub.c =3/2 is non-integral and at each pass the mask will "automatically" fill in the blank spaces left by the previous pass:
______________________________________ pass 1 pass 2 pass 3 pass 4 pass 5 ______________________________________ xoxo oxox xoxo oxox xoxo xoxo oxox oxox xoxo oxox xoxo xoxo oxox oxox xoxo oxox xoxo xoxo oxox oxox xoxo oxox xoxo oxox ______________________________________
(This diagrammatic example symbolizes a real case of, for instance, three passes, a total of ninety-six nozzles used in the pen, thirty-two nozzles used in each of three sections of the pen, thirty-two-nozzle printing-medium advance--and a basic-pattern cell three pixels tall. In algebraic notation, m.sub.p /h.sub.c =32/3, a nonintegral ratio. This three-pass mode is discussed in the next section.)
The print mode produced in this way is essentially a space-rotation mode (though in a sense that condition is not specifically called for). For example, if the pen is a six-row pen as diagrammed above, the first three rows are in "pattern 1" and the second three are in "pattern 2":
______________________________________ xoxo oxox pattern 1 xoxo oxox xoxo pattern 2 oxox ______________________________________
For an autorotating case, either "pattern 1" or "pattern 2" may be used all down the pen. Thus the paper advance turns one simple pattern into a space-rotated mask "automatically". In the shorthand notation introduced above, the pen provides the following periodic behavior as the paper advances.
______________________________________ autorotation ______________________________________ 0 1 2 1 2 1 2 1 2 0 ______________________________________
(g) Three-pass modes--Heretofore, one highly favored print mode has specified a one-third-density-per-pass pattern that constructs dots in a diagonal pattern-- EQU xoo EQU oxo EQU oox
--rather than the one-half-density-per pass checkerboard modes discussed above. The diagonals, however, remain at forty-five degrees as in the checkerboard mode.
This pattern has been considered advantageous because it worked well with software dithering algorithms and had minimal tendency to create moire patterns when printing partial-density-shaded and gradient area fills. The use of forty-five-degree diagonals was considered particularly beneficial for its tendency to distribute error-hiding capability equally between vertical and horizontal axes of the pixel array to be constructed on the printing medium.
Generally a printing apparatus is characterized--through its basic hardware and firmware design architecture --by a general maximum-size print mask or mode pattern that can be formed with the apparatus in one pen pass; any mask pattern to be used with a printing apparatus must fit within its maximum pattern. For example, in a particular one printing device (of the Hewlett Packard Company) which produces high-quality images, that maximum mask or pattern size is eight rows tall and four columns wide--and will readily accommodate, among other possibilities, a mask that is three rows tall (h.sub.c =3) and three columns wide.
Just such a mask produces the one-third-density diagonal three-pass pattern introduced at the beginning of this section. If that mask is used in conjunction with a unit paper advance of thirty-two nozzles--for a printing-medium advance movement m.sub.p =32--then the previously introduced ratio m.sub.p /h.sub.c =32/3, which is not integral.
This combination of conditions accordingly provides autorotation of the three-row mask pattern shown above (as noted parenthetically in the preceding section). No mask rotation sequence is required; and a mask specification for the three passes accordingly might read "111" to indicate that the first column of the base pattern should be used in common to begin each sweep--that is, printing the pixel in column number one of the top row of the swath (assuming that there is any image information to print there). Equally well a mask specification might read "222" or "000", as indeed the pattern may begin with printing any of the three columns of the basic cell.
If instead the number of dot rows were an integral multiple of the pattern height, then as previously explained the printer would have to be instructed to use a rotation sequence telling it how to build the pattern in each succession of sweeps. For example, using the same three-row pattern but thirty-three-nozzle advance--which is to say, a printing-medium-advance movement of thirty-three dot rows--the ratio m.sub.p /h.sub.c =33/3 is integral, and a rotation sequence must be specified.
Such a sequence might be "012"--commanding the printer to form the first swath starting with the first column, number zero ("0") of the base pattern, the second swath starting with column number one ("1") of the base pattern, and the third with column two ("2") of the base pattern, as follows.
______________________________________ pass 1 pass 2 pass 3 ______________________________________ xoo oox oxo oxo xoo oox oox oxo xoo 0 1 2 .rarw. starting column of the base pattern ______________________________________
The other equally acceptable sequences would be "021", "102", and all the other six rotations ("120", "201"; etc.) of these three root sequences. Now if a printer is stopped halfway through a page, using this cell and a diagrammatic six-dot-row paper advance, a pattern something like the following will be found--regardless of whether space or sweep rotation is in use.
______________________________________ xxxxxxxxxxxxxx completely filled xxxxxxxxxxxxxx xxxxxxxxxxxxxx xxxxxxxxxxxxxx xxxxxxxxxxxxxx xxxxxxxxxxxxxx xxoxxoxxoxxoxx two-thirds filled oxxoxxoxxoxxox xoxxoxxoxxoxxo xxoxxoxxoxxoxx oxxoxxoxxoxxox xoxxoxxoxxoxxo oxooxooxooxoox one-third filled ooxooxooxooxoo xooxooxooxooxo oxooxooxooxoox ooxooxooxooxoo xooxooxooxooxo ______________________________________
As before, this abbreviated diagram symbolizes the modernly more interesting practical case of thirty-three-nozzle advance. That case if fully pictured would appear as thirty-three rows fully filled, another thirty-three two-thirds filled, and thirty-three more one-third filled.
(h) Print-quality defects on transparency and glossy stock--As mentioned earlier, known techniques have not been entirely successful in eliminating bulk-colorant problems on transparent and other glossy media. Dividing the total desired amount of ink into three passes has been considered the limit for application of print-mode techniques in attempts to solve this problem.
As noted earlier, evaporation from these media--because they are relatively much less absorbent--is necessarily more important that from plain paper. Some evaporation can be obtained straightforwardly by convection (stimulated by an air-circulating fan), but inducing evaporation by applying radiative heat takes on greater importance with plastic media.
Heat, however, is most straightforwardly applied from below (the opposite direction from that of ink application). These media present more thermal mass and therefore an effectively longer thermal path than does plain paper.
Accordingly with these media a much greater fraction of applied heat radiation ends up absorbed in the printing medium as compared with the ink carrier; this adverse energy distribution is compounded by the previously mentioned dimensional hypersensitivity of these media to heat. Generally speaking, as can be seen from the foregoing discussion, the application of heat is more problematic for glossy and transparent stock than for plain paper.
Heretofore the lower liquid absorption, higher heat absorption, and higher dimensional sensitivity to heating, of these media has defied efforts to obtain adequate liquid removal. Accordingly the prior art has left considerable room for refinement in this area.
(i) Black-ink detail--Printing-machine users often prefer to present lettering and certain other types of finely detailed image elements in black, and the eye is capable of discerning black-inked elements (and defects in them) quite sensitively--as compared with elements and defects marked in other colors. It would therefore be desirable to use finer position control for black inking than for other colors, even within the same image.
Such a strategy, however, is difficult to implement. Generally speaking, the fineness of position control, or to put it another way the pitch of the pixel array, is commonly set by the frequency of a waveform derived by electrooptically reading, while the pen scans, a special scale extended across the printing medium.
Within a printing machine of reasonable cost it is preferable to employ multiplexing techniques for control of the pens. In other words, a single set of signal lines --and control signals time-sharing or otherwise coexisting in those lines--is used to operate all of the pens.
Providing finer position control for printing of black in direct conjunction with other colors would require somehow establishing a separate such waveform for black. That waveform would have to be provided simultaneously with the position-establishing waveform for the other colors--but at a different, higher frequency.
It would also require arranging for the signals of different frequencies to share the same basic position-signal transmitting system. These special provisions, to accommodate established multiplexing arrangements, would be awkward or at least costly. In engineering jargon, electrically it would be hard to "talk" to a color pen (for instance, a cyan pen) and a black pen at the same time.
An alternative would be to print black in a separate sweep, between sweeps for the chromatic-color pens. This alternative would pay a heavy price in reduced throughput and accordingly would be very undesirable.
(j) Conclusion--End-of-page print-quality defects on paper, as well as bulk-colorant problems on glossy and transparent media, heretofore have continued to impede achievement of uniformly excellent inkjet printing--at high throughput--on all industrially important printing media. Awkwardness of overprinting fine detail in black is another adverse limitation of the prior art. Thus important aspects of the technology used in the field of the invention remain amenable to useful refinement.